Alder
Autonomous Molecular Dynamics Research Agent CLI
The Autonomous AI Scientist for Molecular Discovery
Alder is an AI agent that acts as a research assistant, automating the complex coding needed to simulate molecules for drug and materials discovery. R&D teams can now run advanced scientific experiments using simple, natural language prompts.
What is Alder?
Molecular dynamics simulations are one of the most powerful tools in computational chemistry, but running them requires deep expertise in Linux, force fields, and scripting. Alder removes that barrier entirely.
It's a command-line tool powered by a large language model that translates plain-English research questions into fully automated GROMACS simulation pipelines. It plans multi-step experiments, fetches molecular structures, configures force fields, runs simulations, and analyzes results, all from a single prompt.
How it works
From Prompt to Prediction: The Autonomous Agent Loop
User Prompt
Input"Simulate the folding dynamics of Protein X in explicit solvent"
Agent Planning
Custom LLMGenerates a multi-step roadmap with dependency-aware task ordering
Proposal
Human-in-the-LoopPresents planned commands for user approval before execution
Execution
SandboxRuns GROMACS simulations and Python scripts in an isolated environment
Analysis
ResultsParses trajectories, computes observables, and summarizes findings
Current Ecosystem
Tools Alder integrates with out of the box
GROMACS
Industry-standard molecular dynamics engine for atomistic simulations
PubChem
Automated molecular structure retrieval via the PubChem REST API
Python Sandbox
Isolated virtual environment (.alder/venv) with auto-managed Python dependencies
R Workspace
R-friendly project scaffold so Alder can generate and run analysis scripts for downstream statistics and visualization.
Coming Soon
Features planned for the next Alder releases
LAMMPS support
Extend Alder beyond GROMACS with native pipelines for coarse-grained and materials simulations in LAMMPS.
LaTeX report generation
Integrate with a LaTeX Compiler API to turn simulation outputs into high-quality, camera-ready PDF reports.
Web research agent
A literature-aware assistant that scrapes real research papers to surface citations, methods, and references directly into your experiment context.
Remote GPU offloading
Dynamically outsource heavy simulations to a remote GPU compute rig while keeping planning and analysis local.
Key Capabilities
- Natural language interface — describe experiments in plain English
- Autonomous multi-step planning with dependency-aware task ordering
- Human-in-the-loop verification before every execution step
- Sandboxed execution environment with managed virtual environments
- Automated molecular structure retrieval from PubChem
- Trajectory analysis and result summarization built in
Interested in Alder or want to collaborate?